Skip to main content

Full text of "Scaling attractors for quintessence in flat universe with cosmological term"

See other formats

Scaling attractors for quintessence in flat universe with cosmological term 


^Russian State Research Center "Institute for High Energy Physics", 
Pobeda 1, Protvino, Moscow Region, 142281, Russia 
Fax: +7-4967-744937 

^Moscow Institute of Physics and Technology, Institutskii per. 9, Dolgoprudnyi, Moscow Region, 141700, Russia 

For evolution of flat universe, we classify late time and future attractors with scaling behavior 
of scalar field quintessence in the case of potential, which, at definite values of its parameters and 
initial data, corresponds to exact scaling in the presence of cosmological constant. 

PACS numbers: 98.80.Cq, 04.20.Jb 







I • 


Recent astronomical measurements of Super Novae la 
light curves versus their red shifts (SNIa) Cosmic 
Microwave Background Radiation anisotropy (CMBR) 
by Wilkinson Microwave Anisotropy Project (WMAP) 
[4!|, inhomogeneous correlations of baryonic matter by 
Sloan Digital Sky Survey (SDSS) and 2dF Galaxy Red- 
shift Survey with a high precision enforce the following 
picture of cosmology: 

• the Universe is flat, 

• its evolution is consistently driven by cosmological 
constant A and cold dark matter (CDM), that con- 
stitutes the ACDM model. 

Irrespective of dynamical nature for such substances, 
that could be different, at present any model of cosmol- 
ogy has to demonstrate its tiny deviation from the ACDM 
evolution at late times, i.e. the behavior of Hubble con- 
stant should scale extremely close to 

= Hi 


with Ho = H{to) denoting the present day Hubble con- 
stant, a — a{t) is the scale factor in the Friedmann- 
Robertson- Walker metric 

ds2 ^ dt^ - a^{t) [dr^ 

sin'^ Odip'^ 


conveniently normalized by a (to) = 1, so that H — a/a 
with dot meaning the differentiation with respect to time 
t. The fractions J^a and JIm represent the cosmological 
term and pressureless matter including both baryons and 
cold dark matter. In ([1]) we neglect contributions by 
radiation fractions given by photons and neutrinos. 

Dynamical models closely fitting the above behavior of 
Hubble constant include a quintessence , a scalar filed 
(j) with slowly changing potential energy V{(j)) imitating 
the contribution of cosmological constant (see recent re- 
view on the reconstruction of dark energy dynamics in 
[3| ) • In present paper we find a potential of scalar field 
which exactly reproduces the scaling behavior of Hubble 

constant in flat universe with cosmological term^ (Section 
II). The potential is the square of hypersine with some 
tuned values of normalization and slope. In Section III 
we study the stability of scaling behavior versus the pa- 
rameters of potential in terms of autonomous system of 
differential equations possessing critical points. We find 
the late time attractors that further evolve to future at- 
tractors generically different from those of late time. We 
discuss a physical meaning of attractors in Section IV. In 
Conclusion we summarize our results. 


The evolution is described by following equations: 

+ 3i/0-H^ = O, 


where ps is the energy density of baryotropic matter 
with pressure pb = wbPb, which satisfies the energy- 
momentum conservation 

Pb + 3H{pB +Pb) = 0, 
yielding the scaling behavior 

PB = 

qZ(1+wb) ' 



We suppose the following scale dependence of Hubble 

H' = H^ 

,3(i+il.b) y ' 


^ The other approach of reconstructing an effective potential of 
variable cosmological term by the given luminosity distance or 
inhomogeneity growth factor was considered by A.Starobinsky 


where fls denotes the present day fraction of substance 
composed of baryotropic matter, cold dark matter and 
quintessence, that could simulate dark substances ordi- 
nary introduced in the standard consideration: the dark 
energy and dark matter. At late times (or at present) 
wb = corresponds to nonrelativisting matter with neg- 
ligibly small pressure (the dust), while wb = 1/3 stands 
for the radiation era of hot matter. 
The critical density pc is defined by 



so that the baryonic and dark matter fractions are 
presently given by 





while for brevity we put 
^B — + ^DM , 

or pg ^ Pj^+ Pjj^j. 

The scalar field density and pressure 



- V. 

determine the fraction 

^4, = — P0 



and state parameter function 


= - 






The substance fraction is the sum of matter and field 

We define the vacuum energy by 

Vb = Pc^^A, 

so that in flat universe 

flA + ^S ^ 1- 


Since we investigate the scaling behavior of functions 
homogeneous with respect to scale factor a, it is conve- 
nient to introduce the following variable 

N = Ina(t), 


so that the differentiation with respect to time denoted 
by dot is reduced to the differentiation with respect to N 
denoted by prime. 

dV V 


Then, the equation of motion for is deduced to 

^{{cp'Hry + 3{<P'Hr + r ^0. (18) 
The consideration suggests that the potential scales as^ 

V = Vo + 




where fi^ is a constant. Therefore, the derivative of po- 
tential scales, too, 

V = ^3iwB + l)iV- Va). (20) 

According to (flS]) . we suggest 

3{(f>'Hy = -cV, (21) 

where the constant c can be found from ()18|) . since 

{{^'Hfy = -^V" ^c{wB + l)V', (22) 

hence, yields 

c{wB + l)-c+l}V' ^0 

that is satisfied at 

One can easily get 


-{<t>) ^-{<t>H) 

so that 

that gives the relation 

Orf, — 

j3(«;b + 1) I 


1 - Wb 






In order to restore the potential yielding the scaling 
behavior, we have to resolve (|2ip making use of (|19p . 
(El, i.e. 

, 2 2pcf^0 l + WB 1 

' ) ^ 

1-WB a^^'^B+i) ' 


^ Symmetries of evolution equations were systematically investi- 
gated in wherein the authors found the analogous behavior 
of energy density versus the scale factor and used it for deriving 
the equation of state on the basis of supernovae la data, while 
our goal is the potential itself. 



= — Pc{^K + 


The integration straightforwardly yields 

^ « (0 - 0.) = arcsinh^IIlir^, 
where = SttG, and 

A = W 3(1 + wb. 



For brevity of formulae we put the integration constant 
(j>i, = with no lose of generality. Then, 



V = Voil + 




3 1 - uil 
2 A2 ■' 

H"^ = i?o cosh^ i '2 ^ ' 




So, at the present day we have cosh^ K(/>o} = 

Summarizing the result, we emphasize that there is the 
exact solution for the scalar field potential PT|) . which re- 
produce the scaling behavior of Hubble constant in the 
evolution of flat universe in presence of cosmological con- 

The form of potential differs from the case of zero cos- 
mological constant, where the potential is the exponent 
as was studied for the scalar field with the standard ki- 
netic term in [lOj [HI [1.2, 13], while the consideration for 
the general scalar field was developed in [ij, UBI . One 
can easily notice that the present derivation is consistent 
with results concerning for the case of zero cosmological 
constant. Indeed, the integration of (|28p with the Hubble 
rate at f^A = straightforwardly gives the field propor- 
tional to the logarithm of scale factor, cj) ^ In a, that 
makes the scaling behavior of potential V cc exp{XK(f)}. 

The potential derived can be represented in the form 

V = Vo + ^Vo{cosh{XK<l)} 



at Vo = 3Vo{l — w^)/2y. So, function is composed 
by the sum of two exponential potentials with the oppo- 
site slopes and the constant positive shift of minimum. 
Such kind of potentials was investigated recently. 

In review [l6| authors presented the exact solution for 
constant parameter of state for the dark energy in the 

presence of dust-like dark matter, but not adding the 
cosmological constant, which is imitated by the dark en- 
ergy instead. In [l3| the evolution of scalar field with 
the hyper-cosine potential was studied in the case of zero 
cosmological constant: the exponential form found to be 
dominant at early times, while the square term did signif- 
icant at late times. It is clear that the late time dynamics 
essentially changed by the presence of cosmological term, 
of course. 

Various aspects of cosmological picture due to a po- 
tential given by a power of hyper-sine was investigated 
in [l8[ . The tracker properties of such the potentials were 

Another approach was presented in [l^ [l^, HH, [2^ . 
where authors fixed the scaling behavior of Hubble con- 
stant to find exact solutions for the scale factor a{t) in 
order to study characteristics of universe evolution. The 
consideration of recovers the scale factor behavior 
in the case of ACDM, however, the authors did not ad- 
dress the question on the scalar field potential reproduc- 
ing such the scaling. This question was investigated in 
[20], where the potential with the same form of ([M)) was 
deduced in a particular case of VIs/^a = 1/ sinh^ 1 and 
WB = 0. At this choice f^A = sinh^ l/(l-|-sinh^ 1) « 0.58, 
which is in contradiction with the recent measurements 
[j] yielding Oa — 0.766 ± 0.035. A cosmological ex- 
ploration of potential composed by sum of two expo- 
nents with opposite slopes but generically different nor- 
malization factors and a negative shift of minimum was 
investigated in [2l|, by the same method of exact 
time dependence. The authors found an oscillation of 
a{t) around some scaling dependence with oscillating 
within ]-l; -|-1]. 

In [2J] the sum of two exponents with identical nor- 
malization factors but slopes, which can be different, was 
considered at first. The late time behavior of scalar field 
energy scales both the radiation and dust, while near the 
present and future the state parameter has decaying 
vibrations around —1. Such picture differs from that of 
• Questions are the foUowings: i) What is a reason 
for the difference? ii) What can we say about a stabil- 
ity of late time and future scaling? Hi) Does presented 
exact scaling solution corresponds to fine tuned values of 
normalization and slope? These questions were not in- 
vestigated in references mentioned. We address them in 
Section III. 


Let us consider the evolution of flat universe in pres- 
ence of scalar field with potential 

V = Vo + Vo sinh^ <! ^ K, 


where Vq, Vq and A are free parameters, which are not 
fixed by values in (|5D|) . ([?!]) . For definiteness we put 


all parameters to be positive: Vq > 0, Vq > 0, A > 0, 
while the consideration for cases of negative values can 
be rather straightforwardly obtained from the formulae 
below. The Hubble constant is given by 

H' = — l^PB + -{^r + Vo + Vosmh'i^-Kq 

so that we introduce quantaties Uq and [/, so that 


TP — 



Then, the phase space of system is described by dimen- 
sionless variables 

^ = ^~rr^ FT ' ■^ = 77' (37) 

^/6 C/ ' ^ V3 U 
while for convenience we introduce 

1 dV V2 




This choice of variables follows the observation of scaling 
in previous section: the kinetic energy and potential each 
scale like the Hubble constant squared after the subtrac- 
tion of term caused by the cosmlogical constant, while 
the derivative of potential with respect to the field scales 
like the Hubble constant squared itself. 
The definition of implies 

2,2 1 1^'^ PB 

which yields the constraint 



The dynamical state parameter of field is determined by 

P4> + Vq _ 


Pel, - Vb + J/2 

In addition, the equations of motion produce relations 

H = H'H = -{1-wb)pb-{^?, U'U = H'H. (42) 

The differentiation gives the autonomous system of 

x' = — 3a; + ^ 2;z + I X c(a:, y). 

\xz + ^yc{x,y), 


{l + v^)z' = SX^xy ~ ^ zc{x,y), 
v' = ^vc{x,y), 

where c{x, y) = {1 + 'Wb){1 — x^ — y'^) + 2x^ . The quantity 
z is strictly constrained by the condition 

^ /I I 2n 2 Vo 2 


which is the direct consequence of hyper-trigonometry: 
cosh^ q — sinh^ 9=1- This constraint makes system P5)) 
overdefined, since z is completely given by y, w and pa- 
rameter = Vo/Vq. Nevertheless, the system allows us 
to get a complete analysis of critical points in the simplest 
manner. What is of our interest? It is the projection of 
trajectory in the {x,y,v} 3D-space to the 2D-plane of 

A. Late times 

At present, the cosmological constant makes a sig- 
nificant contribution to the Hubble constant, i.e. v — 
Uq/U ~ 1. At late times of evolution just before the 
present, we put v ^ 1. Then, z can be excluded by 


This limit means that the cosmological constant can be 
neglected, while the field has a large value, so that the 
hyper-sine can be approximated by a single exponent. 
Therefore, at late times we arrive to the analysis of ex- 
ponential potentials given in Indeed, under ([^5)1 
system (^5)) is reduced to the system for the exponential 
potential. The analysis of critical point in [llj gave the 
following physically meaningful properties: irrelevant of 
normalization of potential Vq there are stable scaling at- 
tractors in the plane of {x, y}; these attractors appear at 
> 3(1 + wb), so that = AV3(1 + wb) > I 
and W0 ~ Wb- The attractor is the stable node at 
A^ < 24(1 + wbY /{I + 9ws), otherwise it is the stable 
spiral focus. The position of attractors are given by 

3 1 -f Wb 

3(1-^1) 1 


(41) that fixes z according to ([45)1 . i.e. Zc = 

Thus, at late times just before the present, the 
quintessence follows the scaling behavior independently 
of its initial conditions. 

B. Future 

Since function c{x,y) takes positive values at a; 7^ 0, 
?/ ^ 0, i.e. at presence of scalar field, quantity v grows 
in accordance with its differential equation in ()43|) . of 
course. Hence, in future we get v ^ 1. Then, yields 

z ^ Zi, 



i.e., z is frozen at z = z^,. Therefore, we get the system 
for the plane {x,y} in (|43p with z^, being the external 
parameter. The critical points are posed at the following 


I. Scalar field is absent, 

(i): Xi, = 0, 

2/* = 0, 


so that the linearized equations in vicinity of critical 

point, i.e. with x 

with the matrix 



y, result in the 


having eigenvalues 

3{wb - 1) 2* 
-z* 3{wb + 1) 


vi = 





that implies the critical point corresponds to instability 
due to ^tV2 > at wb > 0. Anyway, critical point (l48l) 
is a saddle at < 3 and wb = 0, which is of practical 
interest to look at the present time and future universe. 
We notice that actually the baryonic matter has a small 
pressure, which can be neglected in the universe evolu- 
tion, i.e. Wb +0. At \z^,\ > 3 eigenvalues (fSTI) satisfy 


1^1 = 1^2' ^'^'^ have unstable focus at wb > or center 
at Wb = 0, while at wb < the focus becomes stable. 

II. Critical points of general position are given by 



with additional symmetry over the following permuta- 
tions: A !->■ {Xi, ^ -Xi, and ~y*}, B ^ {x^, ^ y*}, 
C 1-^ {the product of operations A and B}. So, taking into 
account the symmetry, ([5^ covers 4 related sets. The ac- 
tion of permutation A conserves the stability properties, 
while action of B changes them. This fact becomes clear, 
for instance, at < 3. Indeed, altering the sign for 
X implies the interchange of two equivalent branches in 
the potential according to <-^- —(/) or formal altering the 
sign of quantity U, both of which change the sign of y, 
too. So, action A cannot influence the stability prop- 
erties. The action of B interchanges fractions of kinetic 
and potential energies in the energy density, that can be 
physically essential. 

The analysis of linear stability involves matrices for 
([5^ including action of symmetry A and for ([5^ con- 
verted by action of symmetry B, with corresponding 
eigenvalues {v}, 

^ I f WBiV^~zl-3)-2,/9~zl 
^ V -Zi,WB 


-wb{V9^ + 3) - 2^9^ 

■4 = ~iWB - V9^> 

^2 - -v^ 


Bb = 

-u;b(V9+^+3) + 2V9^ 




(V9^-3) + 2y9^ 

-3wb -I- ^9 - zl, 


For real critical points, i.e. at |z^| < 3, and at wb ^ 0, set 
([5^ is stable node (both eigenvalues are negative) , while 
set (|52l) affected by action B gives saddle or unstable node 
depending on the sign of eigenvalue i.e. a balance 
between values of iwB and ^JW 

At \Zi,\ > 3 critical points ([5^ take complex values, 
while matrices satisfy Ba = B^, i.e. they are complex 
conjugate to each other as well as Xi, — y*. The same is 
valid for eigenvectors: 




while e\ 2 — {^12}*- Basis (|55p is orthonormal: 


„a a _ r 

The linearized system has solutions 




where Uj stand for some initial data. Solutions (|56p are 
complex- valued. By our investigation, (j56p are irrelevant 
to physical quantities in question at \zi,\ > 3. 

III. Tuned scaling appears at the special value of pa- 
rameter 2 

Wg. Then, there is the critical line 

(iii): = X* 

1 - Wb 




which is X-invariant, i.e. independent of slope in the 
potential. It is spectacular that tuned point given by 


Zt = Zc, x^, = Xc, Hi, = He corresponds to the exact 
scaling solution found in Section II. 

At (|57p the linear analysis of perturbations gives two 
eigenvalues: vi — and 1/2 — '2wb, so that zero eigen- 
value corresponds to the line itself, while the positive one 
indicates instability at w_b > in vicinity of the line as a 
whole. This fact is in agreement with the study of criti- 
cal points in the case II, since (|57p contains the point of 
(|52| at < as well as it does the point at u;_b > af- 
ter the action of symmetry B, so we correspondingly get 
eigenvalues (j53p and (|54p at z^, z^ yielding the same 
result above. Therefore, at wb > the scaling solution of 
section II is unstable in future, while it does at wb < 0. 

IV. The boundary circle 

(iv): x'^ + 



is conserved by the autonomous system. This fact implies 
that at |z^| > 3 and wb > 0, when there are no stable 
critical points, the system approaches the boundary circle 
in future. 

Summarizing the results, we have found that in the 
presence of cosmological constant the scalar field with 
the hyper-cosine potential approaches the stable attrac- 
tor at late times just before the present day (the time of 
cosmological constant becomes visible) at the potential 
slope > 3(1-1- Wb), so that the energy of field scales 
like that of matter with the state parameter wb, while 
the fraction of field energy depends on the slope value. 
However, this attractor exhibits the strange behavior in 
future, i.e. under the dominance of cosmological term. 
Then, the balance of remnant energy depends on the pa- 
rameter Zi, determined by two quantities: i) the weight of 
potential normalization with respect to the energy den- 
sity due to the cosmological constant and ii) the slope, 
in accordance with (|T7)l . So, at < 3 and wb ^ the 
dynamical part of scalar field energy reaches the other 
scaling attractor, which dominates over the matter in 
accordance with ([55]) (see Fig. [T^) and gets the state 

1 - 



At \zi,\ < 3 and wb < we get an interplay of two 
attractors with ((5T|) and ((53| : at 3wb + ^79 — < 
the scalar field relaxes at the minimum of kinetic and 
potential energy (|48p (see Fig. [T|d), while, otherwise, at 

3wb + — z1 > the scalar field dominates over the 
matter at the same point of ([53|) and (l59|) (see Fig. [TJ:). 

At > 3 we get three cases: i) at wb > the 
quintessence in future vibrates at the boundary circle as 
the limit cycle (see Fig. Wi;), ii) at wb = the field cycled 
around the center being the point of minimal kinetic and 
potential energy (see Fig. [If), ui) a.t wb < the field 
relaxes to the minimum (see Fig. HJi). 

The evolution kinds of scalar field in the plane of {x, y} 
are illustrated in Fig. [T]at A = 20 and various sets of z^, 
and Wb- at z* = 2.9 we put wb = 0.2 in Fig. [T]a), wb = 

—0.5 in b), and wb = —0.2 in c), while at = 10 we 
put Wb — —0.2 in Fig. [T]d), wb = 0.2 in e), and wb = 
in f). At all of tries the evolution starts at xq — —0.8, 
yo = 0.4, and trajectories move clockwise. We certainly 
see that trajectories approaches the late time attractor 
at appropriate {xc,yc}, which are negative in all cases 
except b), when they are positive. The future attractors 
in Fig. [U a) and c) are posed at {a;*,?/*} with opposite, 
negative, sign of (|52|) . The attractors at b) and d) stand 
in the minimum, while limit cycles of e) and f) are at the 
border and around the minimum, correspondingly. 

Thus, in the linear analysis we have classified the late 
time and future attractors for the quintessence with the 
specified kind of potential relevant to the case of nonzero 
cosmological constant. 

However, this analysis falls in special degenerate case 
of z* = 3, wb = 0. 

C. Degenerate case 

At Z-^ — wb — and v ^ 1 the analysis of future 
evolution becomes nonlinear, since, after the transforma- 
tion to variables a — x'^ + y'^ and t = x/y, autonomous 
system (^5)) is reduced to 

, , , 1 - 

cr = 3(7 cr- 1) — 2, 

1 -|- 


which can be solved explicitly. Indeed, the integration 
for T results in 


where 01 = 3(A^— A'o), and Aq corresponds to some initial 
data. Hence, r ^ 1 at the end of evolution, i.e. at 
N +00. Then, 

nnl^^ = hr^^+^,ln(^-01,) 


where cro < Ij = ^(1 + i)- Therefore, at A^ ^ -t-00 
we find 

1 — (T (Tn 

In ^ -21nAf -00, 

1 - CTO cr 

that implies 

cr ^ 1, (63) 
and the attractor is posed at the boundary circle, 



Xi, = y^, = —= or x^, ^ y^, 


If the initial data CTq = 1, the quantity a does not 
evolve, while r approaches the attractor r^, = 1. 


In addition, we have explicitly found that the second 
critical point cr^ ~ 0, i.e. Xi, = y-^ = is unstable. 

The character of attractor ([64|) is illustrated in Fig. [21 
where trajectories move clockwise. 


-0.75 - 0.5 - 0.25 0.25 0.5 0.75 1 

FIG. 2: The attraction of trajectories to the critical points at 
the boundary circle of {x, j/} plane in the degenerate case. 

Thus, attractors (|64p exhibit the stable behavior in ap- 
propriate semicircles, while the line connecting the criti- 
cal points has the instability versus perturbations. 

arbitrary to enlarge the mass up to reasonable values in 
the physics of Standard model. However, huge values of 
Zi, involve extremely frequent oscillations of quintessence 
in the nearest future, that is in contradiction with the 
present smooth evolution of universe. Therefore, wc ex- 
pect that a viable model includes the quintessence mass 
of the order of Hubble constant today. 

B. Restriction to the slope 

The late time scaling of quintessence results in a fixed 
fraction of quintessence energy in the budget of universe 
with respect to other matter irrespective of the evolu- 
tion stage: the dust or radiation fix close values of frac- 
tions. However, the fraction of nonbaryonic matter is 
constrained due to measured and primordial abundances 
of light elements caused by Big Bang Nucleosynthesis 
[Tol . Then, the slope of potential should be quite 

large to suppress fi^ ^ 0.13, so according [ll| one gets 

> 20. (68) 

Next, a role of quintessence field during inflation 
actually was analyzed in pT| , since the hyper-sine in fact 
coincides with the exponential potential at large values 
of field. The problem is a relic abundance of quintessence 
after inflation, that should be small in order to conserve 
the standard scenario of nucleosynthesis. Appropriate 
restrictions in various schemes of inflation are given in 

A. The mass 

The potential of quintessence suggests the mass 

m^ = ^ =47rGA2yo. (65) 
d(j)^ 0=0 

For the exact scaling solution in Section II we get 

ml^^-{l^wl)Hln^, (66) 

and at = and ~ 0.7 of practice the mass is 
determined by the current value of Hubble constant, i.e. 
it is extremely small as well as the energy scale of cos- 
mological constant, that is beyond a natural reason. Al- 
though, such the mass could argue for both the present 
acceleration in the universe expansion and the scale of 
cosmological constant. 
Generically, we get 

2tt 1 
= — VoG zl ^ - H^Qa zl (67) 

So, the mass of scalar quintessence scales as the present 
day Hubble constant with the factor of z^,, which could be 

C. Initial conditions 

Attractors mean a slow dependence of late time evolu- 
tion on initial data for the quintessence. The character 
of regulation is illustrated in Fig. [3l The set of tries 
exhibits the following general features: 

• At small initial fraction of quintessence energy, it 
is frozen to a moment, when it approaches an ap- 
propriate scaling value in order to start the tracker 
behavior at late times. 

• At large initial fraction of quintessence energy, it 
rapidly falls in order to frozen and wait for a mo- 
ment of tracker way at late times. 

• In future, vibrations of quintessence at z^, > 3 and 
wb ^ determine an average value of dynamical 
parameter for the equation of state {w^j,), which is 
independent of initial data, whereas — 1 < (w^) < 
wb at wb > or {w^) = at wb = (see Fig. 
[3] a, b, c, d). At Wb < vibrations determine the 
effective {w^) > wb (see Fig. [Hlg, h). 

Further observations repeat general properties of fu- 
ture attractors. 


-14 -12 -10 -S -6 -4 -2 -14 -12 -10 -S -6 -4 -2 

-25 -20 -15 -10 -5 -25 -20 -15 -10 -5 

-25 -20 -15 -10 -5 -25 -20 -15 -10 -5 

-25 -20 -IS -10 -5 -25 -20 -15 -10 -5 

FIG. 3: Approaching the scaling under different initial data: the logarithm of energy In p in arbitrary units versus the e- folding 
TV = Ina(i) with a casual shift of present day scale factor a{to); the quintessence and baryotropic matter are represented by 
thick solid and dashed lines, correspondingly, while the dynamical part of quintessence without the cosmological term is given 
by thin solid line. Initial data for the suppressed and dominated quintessence are respectively shown in left and right panels: 
a) and b) at A = 20, wb = and 2;* — 50; c) and d) at A = 20, wb = 0.2 and = 10; e) and f) at A = 20, wb ~ 0.2 and 
= 2.9; g) and h) at A = 20, wb = -0.2 and z, = 10. 



-14 -12 -10 -8 -6 -4 -2 

FIG. 4; The state parameter of quintessence versus the 
e-folding A'' of evolution scale: changing the stable value of 
late times at = at = 50 and A = 20. 

D. Equation of state 

At late times, the attractor causes the ratio of 
quintessence pressure to the energy density Wcj, is sta- 
bilized infinitely close to the value of parameter wb for 
the matter. However, the situation changes, when the 
cosmological term comes to dominate, and moves to 
— 1. An example of relaxation is shown in Fig. |4]for 
the quintessence vibrating around the minimum point 
(see Figs. [T]f and [3] a, b). The present day e-folding is 
arbitrary in Fig. [51 The magnitude of deviation from 
the limit of —1 and vibration period depend on the po- 
tential parameters. The picture analogous to Fig. [Hwas 
observed in \2§\ with similar values of parameters. 

It is clear that vibrations are absent if < 3. Any- 
way, the parameter of quintessence state u;^ rapidly ap- 
proaches the vacuum value. Nevertheless, it would be 
interesting to see a relative significance of quintessence 
with respect to the matter. So, general consideration of 
attractors and Fig. [3] demonstrate that the quintessence 
fraction can dominate or be suppressed depending on the 
potential parameters, the value of z^,. It is evident that 
in the case of reaching the boundary circle in the {x, y} 
plane the effective, average value for the state parameter 

is (w^) = 0, while the same value is clearly observed also 
in the case of wb = and \Zi,\ > 3 (see Fig. |3]a and b). 

Thus, we get the definite understanding of phenomeno- 
logical properties for the evolution of quintessence with 
the specified kind of potential in the presence of cosmo- 
logical term. 


In this paper we have found the potential of scalar 
field quintessence, which gives the exact solution for the 
scaling evolution of flat universe in the presence of cos- 
mological constant. The scaling behavior is consistent 
with the current empirical observations. 

We have investigated the stability of scaling behavior 
versus the variations in the slope and normalization of 
potential as well as in initial data. The analysis has re- 
vealed two kinds of attractors. The late time attractor 
just before the cosmological constant is coming to play, 
is independent of normalization, and it is determined by 
the slope, that is consistent with the well-known result 
for the exponential potentials [11], representing the limit 
of large field for the potential found in the paper. The fu- 
ture behavior of quintessence under the dominance of cos- 
mological constant depends on both the ratio of potential 
normalization to the vacuum energy density and slope in 
the special combination denoted by parameter Zi,. Gener- 
ically, the future attractor differs from that of late time. 
So, the late time attractor reveals the strange behavior. 
We have classified the future attractors by their charac- 
ter and stability in linear analysis. The degenerate case 
of nonlinear dependence has been solved explicitly. Some 
phenomenological items have been considered, too. 

We conclude that analysis of scaling attractors can be 
useful for classifying the quintessence behavior at late 
times and in future. 

This work is partially supported by the Russian Foun- 
dation for Basic Research, grant 04-02-17530. 

[1] A. G. Riess et al. [Supernova Search Team 

Collaboration], Astron. J. 116, 1009 (1998) 
farXiv:astro-ph/9805201 ; 

B. P. Schmidt et al. [Supernova Search Team 

Col labo ration], Astrophys. J. 507, 46 (1998) 
arXiv:astro -ph/9805200 ; 

S. Perlmutter et al. [Supernova Cosmology Project 

CoUaborati on], Astrop hys. J. 517, 565 (1999) 
|arXiv:astro-ph/9812133 ; 

J. P. Blakeslee et al. [Supernova Search Team 

Collaboration], Astrophys. J. 589, 693 (2003) 
[arXiv:astro-ph/0302402 ; 

A. G. Riess et al. [Supernova Search Team 

Collaboration], Astrophys. J. 560, 49 (2001) 
[arXiv:astro-ph/0104455 . 

[2] A. G. Riess et al. [Supernova Search Team 

Collaboration], Astrophys. J. 607, 665 (2004) 

arXiv:astro-ph/0402512 . 

[3] P. Astier et al, arXiv:astro-pli/0510447 

[4] D. N. Spergel et al. [WMAP Collaboration], Astrophys. 

J. Suppl. 148, 175 (2003) arXiv:astro-ph/03 02209j ; 

D. N. Spergel et al., arXiv:astro-ph/0603449 
[5] D. J. Eisenstein et al, arXiv:astro-ph/0501171 ; 

S. Cole et al. [The 2dFGRS Collaboration], 

Mon. Not. Roy. Astron. Soc. 362, 505 (2005) 

arXiv;astro-ph/0501174 . 
[6] T. Chiba, Phys. Rev. D 60, 083508 (1999) 

[arXiv:gr-qc /9903094 ; 

N. A. Bahcall, J. P. Ostriker, S. Perlmutter 
and P. J. Steinhardt, Science 284, 1481 (1999) 



arXiv:astro-ph/9906463 ; 

P. J. Steinhardt, L. M. Wang and I. Z latev, Phys. Rev. 
D 59, 123504 (1999) [arXiv:astro-ph/9 812313; ; 
L. M. Wang, R. R. Caldwell, J. P. Ostriker and 
P. J. Steinhardt, _Astropliys. J. 530, 17 (2000) 
arXiv:astro-pli/9901388 . 

V. Sahni and A. Starobinsky, arXiv: astro-ph /06 10026 1 
A. A. Starobinsky, JETP Lett. 68, 757 (1998) 
Pisma Zh. Eksp. Teor. Fiz. 68, 721 (1998)] 






arXiv:astro-ph/98104311 . 


M. Szydlowski, W. Godlowski and R. Wojtak, Gen. 
Grav. 38, 795 (2006) arXiv:astro-ph/0505202 . 
C. Wetterich, Nucl. Phys. B 302, 668 (1988). 
E. J. Copeland, A. R. Liddle and D. Wands, Phys. Rev. 

D 57, 4686 (1998) [arXiv:gr-qc/97 11068 . 

P. G. Ferreira a nd M. Joyce, Phys. Rev. D 58, 023503 

(1998) [arXiv:astro-ph/9711102; . 

A. Albrecht and C. Skordis, Phys. Rev. Lett. 84, 2076 
(2000) arXiv:astro-ph/9908085 . 
S. Tsujikawa, arXiv:hep-th/0601178 

Y. Gong, A. Wang and Y. Z. Zhang, Phys. Lett. B 636, 




286 (2006) arXiv:gr-qc/0603050 . 

V. Sahni and A. A. Starobinsky, Int. J. Mod. Phys. D 9, 
373 (2000) arXiv:astro-ph/9904398 . 
V. Sahni and L. M. Wang, Phys. Rev. D 62, 103517 
(2000) arXiv:astro-ph/9910097 . 

L. A. Urena-Lopez and T. Matos, Phys. Rev. D 62, 
081302 (2000) arXiv:astro-ph/0003364 .' 
A. Gruppuso and F. Fin eUi, Ph ys. Rev. D 73, 023512 
(2006) arXiv:astro-ph/051264T 1 . 

A. A. Sen and S. Set hi, Phys. Lett. B 532, 159 (2002) 
arXiv:gr-qc/011 1082| . 
C. Rubano, P^ Scudellaro, E. 
S. Capozziello, Phys. Rev. D 
arXiv:astro-ph/0311535 . 
C. Rubano^ R Scudellaro, 
S. Capozziello and M. Capone, 

Piedipalumbo and 
68, 123501 (2003) 

E. Piedipalumbo, 
Phy s. Rev. D 69, 
103510 (2004) |arXiv:astro-ph/0311537]. 
T. Barreiro, E. J. Copeland and N. J. ISJunes, Phys. Rev. 
D 61, 127301 (2000) arXiv:astro-ph/9910214, .